Something like SIMPER analysis in Primer-E?

[Elizabeth Bent]

Hi, I'd like to use PC-ORD data to figure out something like Primer-E's SIMPER analysis. The SIMPER analysis identifies taxa most responsible for the multivariate distance between experimental treatments (or samples, if this is too difficult).

I was thinking I might use the distance of taxa points (column data) from the origin of my NMS plots to estimate taxa that are most responsible for the observed Bray-Curtis distances between my samples, but I'm not sure this is appropriate, and if it's possible to identify taxa that cause differences between treatments, I'd like to do that. 

I'm not much of an ecologist, so I'm learning as I go. I don't want to do an inappropriate analysis.

Liz

[Bruce McCune]

Liz,
You could do something like that, but it wouldn't give you a hypothesis test and might not be as informative as some other approaches. The most commonly used approach is to use indicator species analysis for each taxon, to complement an overall comparison for the null hypothesis of no difference in communities between treatments. The latter could be done with perMANOVA or MRPP or SumF.

Good luck,
Bruce McCune

--
You received this message because you are subscribed to the Google Groups "PC-ORD" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pc-ord+un...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Elizabeth Bent]

Hi Bruce- thanks! I've already done MRPP tests so I know there are differences between my treatments, and I'm only interested in ranking taxa in terms of how much effect they have on the distances between my samples. I will take a look at indicator species analysis as well. How does an indicator species analysis differ from a SIMPER-like analysis?

Liz

--
You received this message because you are subscribed to a topic in the Google Groups "PC-ORD" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/pc-ord/fUNyh2U58Ag/unsubscribe.
To unsubscribe from this group and all its topics, send an email to pc-ord+un...@googlegroups.com.

[Bruce McCune]

Liz, SIMPER and ISA are different mathematically, but that's not to say that SIMPER and ISA wouldn't give similar results. SIMPER is based on partial similarities (individual species contributions), while ISA has its own metric for differences in individual species between groups, a metric that is quite logical for species data. I don't know of a formal comparison.

Bruce McCune